School Matters

By Peter Mortimore, Pamela Sammons, Louise Stoll and 2 more

This title is part of UC Press's Voices Revived program, which commemorates University of California Press’s mission to seek out and cultivate the brightest minds and give them voice, reach, and impact. Drawing on a backlist dating to 1893, Voices Revived makes high-quality, peer-reviewed scholarship accessible once again using print-on-demand technology. This title was originally published in 1988.

• Language: English
• Publisher: University of California Press
• Release date: Apr 28, 2023
• ISBN: 9780520330375

Peter Mortimore

Enter the Author Bio(s) here.

Book preview

School Matters

School Matters

Peter Mortimore, Pamela Sammons, Louise Stoll,

David Lewis and Russell Ecob

The University of California Press

First published in 1988 by the University of California Press

© Peter Mortimore, Pamela Sammons, Louise Stoll, David Lewis and Russell Ecob 1988

All rights reserved. No part of this publication may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopy, recording, or any information storage and retrieval system, without permission in writing from the publisher.

Library of Congress Cataloging-in-Publication Data

School matters.

Bibliography: p.

Includes index.

1. Education, Elementary — England — London — Evaluation — Longitudinal studies. I. Mortimore, Peter.

LA633.S35 1988 372.9421'2 88-17299

ISBN 0-520-06502-6

ISBN 0-520-06503-4 (pbk.)

lÿpeset by Character Graphics,Taunton

Printed and bound in Great Britain by

A. Wheaton & Co Ltd, Exeter

Contents 1

Contents 1

1 The Search for Effectiveness

2 The Schools

3 The Headteachers and Deputies

4 The Teachers in the Classroom

5 The Curriculum

6 Pupils’ Progress

7 Different Groups

8 Teacher Expectations

9 The Importance of School

10 Differences in Effectiveness

11 Understanding Effectiveness

12 Towards More Effective Schooling

References

APPENDIX 1: The Secondary Transfer Project

Index

1

The Search for Effectiveness

WHY HAVE A STUDY OF JUNIOR SCHOOLS?

Does the particular school attended by a child make a difference? Will a child’s progress in reading or writing be similar wherever she or he is taught? Are some schools more effective than others? These are just three of the questions, often posed by parents, which have recently stimulated researchers to investigate how much difference there is between the most and the least effective schools. Curiously, traditional educational research, carried out in Britain and in North America, has failed to address these important questions. This was probably because social scientists, such as Coleman et al (1966) and Jencks et al (1972), had argued that home background, including social class and economic status, were much more influential on a child’s development. They reasoned that because the differences between families were much greater than those between schools, families were likely to exert the greater influence. Whilst it is undoubtedly true that an economically advantaged family — with comfortable housing, healthy diet, and time for stimulating educational experiences, contrasts starkly with an economically disadvantaged one — with inadequate, overcrowded or even a lack of permanent housing, poor diet and little time or money for educational experiences, it is also true that schools vary a great deal. The problem for researchers is how to tease out the effects of families from the effects of schools.

Two groups of research studies — one in the United States and one in Britain — have recently begun to address this issue. In the United States, a group of researchers including Weber (1971), Summers & Wolfe (1977), Brookover et al (1978,1979), Edmonds (1979a, 1979b), Good (1979), and Goodlad (1979), have been active in challenging the view that the influence of school can only be trivial. In Britain a group of researchers (see Reynolds, 1985) concerned with school differences has carried out a number of pioneering studies designed to separate out the effects of home and school. Notable amongst the British group have been a study of Welsh secondary schools (Reynolds, 1982) and a study of 12 inner city secondary schools (Rutter et al, 1979). This latter study — written up in a book entitled Fifteen Thousand Hours; Secondary Schools and their Effects on Children — was important because it developed a methodology to evaluate the effectiveness of schools after having taken account of the characteristics of the pupils entering those schools. The research team involved in Fifteeen Thousand Hours included one of us.

Although these studies attracted methodological criticisms, their impact was to encourage other researchers to develop suitable techniques (see Aitkin et al, 1981a, 1981b; Gray, 1981a) and appropriate methods of analysis (Goldstein, 1980,1984) to address this important question in a more sophisticated way. A recent American review of the area of school effectiveness (Purkey & Smith, 1983) demonstrated that these studies have succeeded in identifying a number of common factors. In other words, we researchers have managed to clear the decks, and to develop generally accepted ways of answering the questions with which we began this chapter.

Because most of the British research on school effectiveness has been carried out in secondary schools, we decided to focus on primary education. Somewhat surprisingly, this area of schooling had been rather neglected. With the exception of Bennett’s (1976) study of teaching style, the 1978 survey by Her Majesty’s Inspectors, the study of teachers carried out by researchers from Leicester University (Galton et al, 1980; Galton & Simon, 1980; Simon, 1980) and a number of smaller-scale projects (for example, Armstrong, 1980), education for pupils below the age of 11 at the beginning of our study was remarkably unresearched.

We decided, therefore, to fill this gap by mounting a study of the junior years of primary schooling. In order to be able to examine the progress of pupils, we knew the study would have to be longitudinal and follow pupils through several years of school life. We also knew that if we really wanted to discover whether individual schools made a difference our work would have to be very detailed. Schools are complicated places — as many parents have discovered. Finally, we knew that if we wished our work to be taken seriously by parents, practitioners and researchers it would have to be very rigorous, demonstrate that the lessons of earlier work had been learned and employ the appropriate methods of analysis, however sophisticated these were.

What follows is our account of the study. The bare bones are that we followed a group of 2,000 pupils through four years of classroom life (from age 7 to age 11) in 50 schools selected randomly from the 636 schools in the Authority. In order to put flesh on these bones, however, we spent the four years in close observation of the classrooms and the schools through which our sample of pupils were passing. This book is the outcome of the study: an account of the progress of the pupils, and a description of their school days. We start with a formal statement of our aims.

The aims of the study

We had four main aims: The first was to produce a detailed description of pupils and teachers, and of the organisation and curriculum of the schools. The second was to document the progress and development, over four years of schooling, of nearly 2000 pupils. Our third and key aim was to establish whether some schools were more effective than others in promoting pupils’ learning and development, once account had been taken of variations in the characteristics of pupils in the intakes to schools. The fourth was to investigate differences in the progress of different groups of pupils. Special attention was paid tc achievement related to the race, sex and social class backgrounds oi pupils. In addition, we wanted to examine the effects of differences of age on children’s achievement. In order to pursue these aims, we addressed the following questions: Are some schools or classes more effective than others, when variations in the intakes of pupils are taken into account? Are some schools or classes more effective for particular groups of children? If some schools or classes are more effective than others, what factors contribute to these positive effects?

METHODS USED IN THE STUDY

The data

Our data can be divided into three categories: measures of the pupil intakes to schools and classes; measures of pupils’ educational outcomes; measures of the classroom and school environment.

1) ) Measures of the pupil intakes to schools and classes

We collected detailed information about pupils’ characteristics in order to explore the effectiveness of schooling for different groups of children (according to age, social class, sex or race), and also to explore the impact of background factors upon educational outcomes.

Previous studies of school effects have been criticised on the grounds that their measures of intake were inadequate, and because they collected data at the level of the school, rather than the individual pupil (see, for example, work by Marks et al, 1983, which was criticised because of its failure to take proper account of differences in intakes, by Gray, 1983, and Gray and Jones, 1983). This sounds a technical and not very important point, but it is, in fact, crucial for all studies of school effectiveness. Whereas some researchers have related the average achievement of pupils in a particular school to the proportions of parents coming from particular social classes (this being better than ignoring social background altogether) we were able to relate the achievement of individual pupils to their own characteristics, including their particular social class background. Because of the use of individual data, it was possible for us to use more sophisticated techniques of analysis and, therefore, to take a fuller account of the impact of background factors than has been the case in many previous studies.

The background measures of pupils’ characteristics used in our study covered two areas: the social, ethnic, language and family background of the children, and their initial attainment at entry to junior school. All information was obtained at the level of the individual child and, because of the longitudinal nature of the study, it has been possible for us to explore the cumulative effects of background factors over several years.

We also collected data about each child’s attainments in assessments of reading, mathematics and visio-spatial skills, and obtained a class teacher’s rating of behaviour at entry to junior school. This information enabled account to be taken of differences in the past achievements and development of pupils, which may have been influenced by their previous infant classes and nursery schools. It also provided the necessary baseline against which assessments of the later progress and development of individual children, during the junior school years, could be made.

2) Measures of educational outcomes

The results of studies of school effectiveness are dependent, to a large extent, on the choice of measures of educational outcomes. Most studies of secondary school effects have been criticised for concentrating on too few measures of educational outcomes (usually examination success and attendance), and studies of the junior age group have frequently focused only on children’s attainments in the basic skills.

In our view the aims of primary education are, rightly, diverse. Basic skills are considered important by the vast majority of teachers and parents, but other areas — including aspects of development such as behaviour, attendance and attitude towards education — are also important. Studies which use only one or two measures of pupils’ educational outcomes can give an unbalanced and simplistic view of class and school effects.

a) Cognitive outcomes We employed standardised tests of reading and mathematics. The tests used were: the Edinburgh Reading Test (ERT) and the National Foundation for Educational Research Basic Mathematics Test (BMT). These tests have been shown to be reliable and to be reasonable predictors of later academic success. Because of the considerable variations in children’s attainments in these skill areas at entry to junior school, pupils were assessed regularly to enable us to investigate their progress over the junior years.

In addition to these tests we devised individually-based assessments of practical mathematics which were conducted in each school year. Also, to take account of the importance of writing in the junior cur riculum, an assessment of creative writing was made on an annual basis. Furthermore, because of the acknowledged importance of pupil talk and interaction to their development in school, the oral skills of a sample of children were assessed in the fourth year, using exercises developed specifically for the study by the Language Survey Team of the Assessment of Performance Unit (APU) of the Department of Education and Science (see Gorman and Hargreaves, 1985).

Finally, during the fourth year we also collected information about each pupil’s performance in the Authority-wide pre-secondary transfer tests of reading — the London Reading Test (LRT) — and verbal reasoning.

b) Non-cognitive outcomes We collected information about the children’s behaviour in school (as assessed by their class teachers) using an instrument specially developed for the Project. This information was collected for each child in the autumn and summer terms of each school year. In this way it was possible for us to examine changes in behaviour over the years. In addition, a self-report measure of pupils’ attitudes towards different types of school activities, curriculum areas and other aspects of school was used in each school year. Measures of each child’s perception of how they were seen by the teacher and by their peer group, as well as their views of themselves, were obtained at the end of the third year, and full attendance data were collected about all children, individually, in each of the three terms of the first, second and third years.

3) Measures of the classroom and school environment

Our third group of data relate to the teachers, classrooms and schools involved in the study. Because of the interest in identifying which factors make some schools or classes more effective than others in promoting good educational outcomes, a wide variety of information about the teachers, classes and schools was collected.

a) School organisation and policies Information about school organisation and policies was obtained through interviews with heads, and deputy heads. Questions concerning their role, educational philosophy, qualifications and experience were included. Particular attention was paid to the way pupils were grouped into classes, the allocation of teachers to classes, the allocation of staff responsibilities, and teacher involvement in decision-making.

b) Class organisation and policies The class teachers of pupils in the sample were questioned about their qualifications, responsibilities, philosophy of education and involvement in decision-making. Information about their methods of assigning work to pupils, and their grouping strategies within the class, was also gathered. A plan of each classroom was obtained so that seating arrangements and layout could be examined. Detailed information about special needs teaching, the curriculum, and use of timetables was also collected.

c) Teacher strategies Classroom observations were undertaken in each of the three years. A systematic procedure was adopted because of the necessity for comparability over time and between classes. The instruments developed by the ‘ORACLE’ team from Leicester University were chosen because these had been used recently with a similar age group, and because they provided useful information about teachers’ and pupils’ activities (see Boydell, 1974, 1975; and Galton et al, 1980). In addition to the ORACLE system, an amended version of the ‘SCOTS’ schedule (Powell & Scrimgeour, 1977) was used to provide more subjective ratings of classroom behaviour and the activities of teachers and pupils. Extensive use was also made of qualitative data, including notes, case studies and verbatim descriptions provided by the field officers.

d) Views of parents Parents of all the sample pupils in eight of the schools were interviewed. These interviews, which took place in the child’s home, were carried out by specially trained interviewers matched for ethnic background and speaking the appropriate home language. Parents provided validation of information obtained from schools about their child’s background characteristics, and were asked about their views of their child’s progress, and their own involvement with her or his learning.

e) School life We used a wide variety of methods to collect information about the pupils, classes, teaching staff and schools involved in the study. These included assessments of pupils (using written and practical tasks), self-report questionnaires and teachers’ assessments of behaviour. The latter were used for the analysis of pupil progress and development. The use of questionnaires, interviews and observational data enabled a check on the validity of different measures to be made. This is important, for social psychology has shown that people are not always fully aware of their own behaviour. An interesting example of this occurred when the majority of teachers reported in interviews that they spent most of their time dealing with the class as a whole, rather than with individuals or groups. From systematic observations, however, it was clear that, in every year, teachers spent much more time communicating with individual children than with either the class as a whole or, even more extremely, with groups. These results, suggesting that peoples’ perceptions of their own behaviour may not always correspond with what they actually do, point to the necessity of obtaining information from a variety of sources.

METHODS OF ANALYSIS

Analyses of our data were necessarily complex. Early research on school and teacher differences was criticised on methodological grounds (see for examples of the debate, Mortimore, 1979; Tizard et al, 1980; Goldstein, 1980; Rutter et al, 1980; Radical Statistics Education Group, 1982). The implications of such criticisms for analyses of school and teacher differences have been illustrated by re-analyses of the Bennett study which have resulted in substantial revisions to the original findings (see Bennett, 1978; Aitkin et al, 1981a; Prais, 1983; Chatfield, 1985).

As with all field work carried out in schools, the research design had to accommodate severe limitations. In order to overcome some of the methodological problems caused by these limitations, a variety of different statistical techniques were employed. The use of measures developed from quite different methodologies enabled checks on validity to be made. Whenever possible, test scores were adjusted for unreliability, and account was taken of the impact of clustering within the sample of pupils at the level of the school.

Three crucial aspects of the effects of junior schools on their pupils were investigated in detail.

The size of the effects, in terms of the proportion of the overall variation in pupils’ progress or development which can be explained (in a statistical sense) by different schools, in comparison to that explained by the children’s background characteristics.

The size of the effects of individual schools on their pupils’ educational outcomes.

(For example, what difference in a pupil’s reading progress over three years can be attributed to her or his membership of a particular school?)

The processes which relate to the effects of individual schools and classes on their pupils’ educational outcomes.

In examining the overall impact of school membership during the junior school period, it was possible to draw on data collected over three full years. When the size of the effects of school and class membership were compared, however, it was necessary to examine the data for each year separately. This was because of the frequency of changes of teacher and pupil membership of classes between years. The use of a multilevel model for much of the analysis has enabled the separate effects of school and class, over the period of one school year, to be identified and studied (see Goldstein, 1986).

In addition to analyses of school and class effects, the attainments, progress and non-cognitive development of children were examined for all individuals and, separately, for different groups. When investigating differences in outcomes due to age, social class, sex or race, the analyses controlled simultaneously for all other background factors. This means that we could isolate the separate effects of any given factor, when the combined influences of all other background characteristics were taken into account. Thus, the effect of sex, for example, was identified net of the effects of age, social class, race and other background factors.

In addition to examining the overall relationships between progress and achievement in different cognitive and non-cognitive areas for all pupils, it was possible to examine the relationships for children with different characteristics (according to age, social class, sex and race). The relationships between attainment and progress and teachers’ ratings of pupils’ abilities were also investigated. Again, the analyses were conducted for all pupils and, separately, for children of different groups.

This was our research strategy in 1980. Seven years later, having collected an enormous amount of data about pupils, teachers and schools, subjected these data to rigorous statistical analysis, and attempted to interpret our findings in the light of the results of other research and our own knowledge of schools, we are ready to report our conclusions. The following 11 chapters represent our attempt to do this. We have tried to write in a non-technical way (and have published separately information on technical issues) but we hope we have included sufficient information about our data for readers to make their own judgements of the findings, and thus not to be restricted to our interpretations.

We have also included a considerable amount of description of the classes and schools which were observed in the course of the Project. Whilst not essential to our central argument about the impact of schools, we believe this material will be of considerable interest to both teachers and parents. In the light of the lack of detailed descriptions of what schooling is like during the years of childhood, we hope our contribution will prove useful.

In the following chapters we deal in turn with the schools; the teachers; the classroom; the curriculum; the pupils and their attainments, progress and development; and the expectations of the teachers, before turning to the results of analyses designed to test whether individual schools had differential impacts upon their pupils’ educational outcomes. The final chapters examine the factors which contribute to effectiveness, and discuss some of the implications of our findings.

2

The Schools

Over the course of the Project, much information was collected about the 50 schools in the sample. This information covered both the givens, as we labelled them, and the policies. The givens covered all aspects of structure and organisation outside the direct control of the headteacher and the rest of the staff. These include: the school building itself; resources dependant upon external funding; status, whether with (JMI) or without infant departments (JM); whether county or voluntary; and the size of the school, in terms of pupil roll. The policies incorporated any features over which the head and staff could exercise control. Policies comprised, amongst others: the curriculum; arrangements for allocation of pupils, and teachers, to classes; schools’ promotion of equal opportunities policies; and the involvement of parents in the life of the schools.

SCHOOL ‘GIVENS*

School buildings

Some schools were built in the latter half of the last century and have been partly or completely reconstructed. Others are housed in the original Victorian ‘three-deckers’ which have become the hallmark of the typical London school. These old buildings, however, have often been extended and adapted to accommodate more recent educational practices. The separate entrances for girls and boys, for example, are now merely a reminder of Victorian attitudes. Nearly half the schools were Victorian ‘three-deckers’. The next largest group were buildings from the 1950’s and 1960’s. The rest had been built between 1970 and the present day.

A particular feature of schools in the inner city is that their space is limited. Few of the schools in the survey, for example, had playing fields. Most, however, had a small area set aside for plants, and a few also had a grassy play area. Nearly half the schools had two playgrounds, and these were sometimes brightened by murals.

Use of rooms within the buildings

Space was seldom so short that any class was without its own room. In one school, however, two classes did have to share a room and each class spent part of their day in a small open area in the corridor. Six schools had some open plan classrooms though, in a few cases, separate areas had been screened off or the entrance was bricked up.

All but seven of the schools had a library. Some had separate libraries for infant and junior children; others for fiction and non-fiction books. Libraries were also commonly used for other activities. Examples of these included use as a pupils’ resource section, storage for audio-visual aids and space for watching television. Libraries were also often used as a place for withdrawal of small groups who needed extra help.

Due to a shortage of space in many schools, accommodation used for specialist purposes had to be shared. For example, sometimes mathematics and language areas were combined. Similarly, in certain schools, areas were set aside for music and art, or science and general resources. The specialist area that most frequently had its own space, however, was special needs teaching.

In 11 of the schools the importance of co-operation with parents was emphasised by having a room specially designated for their use. This gave parents the opportunity to meet informally, as well as providing them with a space within the school for classes in English or, in some cases, handicrafts.

Accommodation difficulties

Inevitably, some buildings posed problems. When interviewed, the majority of teachers cited difficulties, most of which were concerned with the provision of facilities, space or layout. One example was the inconvenience of teaching in a hall, when this also formed a thoroughfare for teachers and pupils. Some teachers also complained about difficulties of managing pupil movement and noise on the stairways in old buildings. Other problems arose from the necessity to remove asbestos, lack of display or storage space, or were a result of having to share facilities.

Resources

Resources for most curriculum areas, in the main, were available within all schools. Provision of library books, on the whole, was good, particularly non-fiction material. Art materials were usually plentiful, and there was a variety of musical instruments in most schools. Drama equipment, however, including stage boxes and lights, was much less frequently seen. Practical mathematics resources, such as scales, blocks and clocks, were in good supply in all but three schools. During the course of the Project, the use of computers in schools became increasingly widespread; one school had eleven. There was also a fairly wide selection of audio-visual equipment in most schools.

The visitor’s view of the schools

Even before they meet any pupils or teachers, visitors can gain an impression of the school from walking in the corridors and open areas. Quite apart from the general state of repair, their feelings may be influenced by the nature and extent of displays of work, notices welcoming parents into the school, and even the presence of plants or flowers.

Schools were rated by us on their decorative order. Two thirds were in good condition, and most of the rest were rated as being fair. A very similar impression was gained of the condition of furniture. Graffiti on walls and tables was only noticed in seven schools, and even then it was not extensive. Decorators were seen in several schools and building work also took place in a few. This, however, sometimes caused considerable upheaval.

Just over half the schools had a noticeboard for parents and visitors. Generally, messages were written in English, but in three schools all notes were also translated. Many schools displayed a multi-lingual ‘WELCOME’ poster.

Schools also varied considerably in the amount and type of displays seen in corridors, on stairways and in the hall. Although some schools used all or nearly all available space for showing work, a few others made no use of it. In two schools, even the display boards were empty. Overall, in six schools, work demonstrating a wide selection of art and craft techniques was on show. However, in more than a third of schools, given the availability of resources, the range was surprisingly limited.

Photographs of school plays, journeys, clubs, teams or children at work were frequently mounted and displayed. Notices about school clubs and teams were also common. In a few schools interest tables, such as ‘the seashore’, were observed.

In most schools, pupil work reflected all ages and abilities. However, in a few it was clear that some teachers did not display the work of their pupils, and in some, only the best work was exhibited. In others, every pupil had an equal opportunity of seeing her or his work exhibited around the school.

Thus, although all of the 50 schools can be described as inner city schools, they displayed marked differences both in their physical structure and the uses to which rooms were put. They also varied considerably in the impression created by displays in open areas.

SCHOOL ORGANISATION

Schools with and without an infant department

In their interviews the heads of the 36 schools with infants described to us their advantages. The majority stressed continuity of the pupils’

S.M.—B education. One head reported: ‘The children know the staff and the building, so they only have to cope with junior life’.

Most headteachers also felt that the staff benefitted from the broader age range in the school. In this way they could get to know the pupils better, and make use of their knowledge, whilst broadening their teaching experience by working with both infant and junior classes.

A third of the heads also talked of the advantages of the combined system to themselves, and the broadening of their own experience. Closer involvement with parents was mentioned by several others.

One headteacher summed up all the positive comments when he said: ‘I am convinced that the primary school as a totally embracing unit is the best possible arrangement for educating children’.

Two-thirds of these heads could see few disadvantages to their system. The small number of problems that were noted tended to relate to the headteacher’s own lack of experience with either the infant or junior age range, or to extra organisational requirements.

For heads of those schools without infant departments, however, concern was expressed about the different ideas, philosophy and style of the junior and infant headteachers. As one head said: ‘A child moving from one school to another, even in the same building, finds different patterns of organisation and outlook. Initially there are some problems connected with settling in. These would be obviated if they went to only one school*. Only one headteacher felt strongly that the separate system had considerable advantages, and that was only if a child had been labelled in her or his infant years and needed a fresh start.

Although headteachers were almost unanimous in favouring schools covering the primary age range, some class teachers of first year pupils nevertheless felt that there was a difference in teaching style and philosophy between the infant and junior departments of the same school.

County and voluntary-aided schools

Of the 50 schools initially involved in the Project, 35 were maintained fully by the Authority. Eight of the rest were voluntary Church of England schools, six were Roman Catholic schools, and one was managed in association with a non-denominational body. The ratio of the county to voluntary schools in the Project sample was similar to that in the Authority as a whole. All of the voluntary-aided schools covered the whole primary age range, whereas more than a third of the county schools only contained junior-age pupils.

Difference in the size of the sample schools

First year roll

The 50 schools differed markedly in the size of their first year junior intakes. The largest school received a total of 102 pupils at entry, whereas only 16 entered the smallest school. This compares with an average first year roll of 36 pupils for the 50 schools as a whole. In all, ten schools had first year rolls of 45 or more pupils and ten had fewer than twenty-six.

Previous studies (Plowden, 1967; Barker Lunn, 1982) have suggested that schools with a small intake or overall roll may well be restricted to certain types of organisational strategies because of their size. Variations in strategy according to school size will be described later.

Overall junior pupil rolls

There were considerable differences between the schools in terms of overall junior rolls. Thus, at one extreme, 519 pupils were on roll at the beginning of the academic year 1980, compared with only 73 in the smallest school. The average junior roll for the 50 schools as a whole was 159 children. In all, ten schools had a roll in excess of 200, whilst at the other extreme, the smallest ten schools had rolls of fewer than 112 junior pupils. In the Authority as a whole at the same time (September 1980), a third of the JM and JMI schools had less than 112 junior pupils on roll. Conversely, 23 per cent had 200 or more junior pupils on roll. Such differences in overall school size have an impact on staffing levels and may influence organisational policy and the availability of posts of responsibility (see Luzio, 1983).

SCHOOL POLICIES

It has already been shown that schools varied considerably in their junior roll. We now turn to the policies affected by the size of the roll, both in terms of numbers of pupils and availability of staff.

Number of classes and class size

The average number of junior classes in both JM and JMI schools was eight, although there were five schools with less than five classes and five schools with 12 or more.

Although the planned pupil-staffing ratio of the schools was based on the roll, some schools had additional posts for special needs. Additionally, some heads had used their own resources (funded through a special Authority scheme) to pay for other posts. As a result of these factors, class sizes often differed considerably. The numbers of pupils in different years also often varied and a year group, in some schools, could be split into two unevenly sized classes. Sometimes this was because particular pupils were perceived as needing more teacher attention. In other cases, a particular teacher (often a new member of staff) was given a smaller class. In three schools teachers worked in teams. The mixing of pupils of different ages was also fairly common.

Some headteachers took the view that such a policy allowed greater flexibility of class size although, as is discussed later, others suggested that it could pose problems.

Over three years a total of 278 junior classes contained pupils in our sample. The average overall class size was 25. This figure was based on the actual numbers of pupils in each of the classes we surveyed in the Project. The average figure is somewhat higher than those calculated annually by the Inner London Education Authority (ILEA). This is because the latter take into account all teachers who are involved with class teaching. This includes nurture groups and those teaching staff (such as the head and deputy) who may teach a number of sessions, but are not actually responsible for a class. Thirty-three classes contained more than 30 pupils, though in nine of these team teaching was in operation. In 18 classes, there were 17 or fewer children. Within one school, the variation between two classes was 15 (one class contained 27 pupils; in another class there were only 12).

Criteria for allocation to class

In allocating pupils to a particular class the headteachers had to take two decisions: whether to keep children of the same age together in single-age classes or to mix age groups; and whether to take account of other factors relating to the pupils, such as their ability, behaviour or sex.

Single or mixed-age classes

In 22 of the schools, all pupils were allocated entirely on a year group basis. In the other schools, pupils either were organised into vertically grouped classes, or into a mixture of single and vertically grouped classes.

Headteachers were questioned about the reasons for adopting particular systems. In nearly all schools, falling rolls had influenced the decision to have mixed-age classes. A mixed-age system had been introduced, solely for educational reasons, in only two schools. These headteachers saw benefits such as security and continuity which, in their view, also allowed able pupils to be stretched.

Concern was expressed, however, by nine headteachers who thought that mixed-age teaching was more difficult for class teachers to organise. This was a view shared by Thomas (1985). Other disadvantages noted concerned the problem of coping with a two year age range, in addition to the wide ability range found within a single year group. Some heads also felt that more able children might not be sufficiently stretched if the work was geared towards the middle of the ability/age range.

Other criteria for the allocation of pupils to classes

Approximately nine per cent of heads grouped children according to their stage of learning. In almost all cases, in those schools where a mixed-age system was